Just in case you’re interested in the way we feel about the people who trust us with their investment capital, I wanted to share… In December 2024, we bought back 11.76 million shares of $clsk for $145M. In November 2025, we bought back 30.6 million shares of $clsk for $460M. We extinguished all equity offerings. We shut off our atm. We’ve stacked 13,000 Bitcoin—all mined. We’ve closed on more land. We’ve secured more power. We’ve started assembling a world-class AI data center team. We bet on us. We now control a significant portion of the most scarce assets on the planet. $btc $clsk and powered land. @CleanSpark_Inc Thank you for taking this journey with us.

Today $CLSK reported transformative full fiscal year 2025 financial results (ended 9/30/25), setting the stage for our AI expansion. *Revenue: $766.3 million (102% growth YoY) *Net Income: $364.5 million *Total Assets: $3.2 billion *Power Under Contract: 1.31 GW (as of 10/31/25) Full press release here: clsk.news/fy25results “Fiscal 2025 was the year CleanSpark achieved operating leverage. We surpassed 50 EH/s in operational hashrate, set new revenue records, and demonstrated strategic capital stewardship by choosing accretive capital market tools, such as convertible debt and bitcoin backed revolvers instead of an ATM to finance the business during the calendar year,” said @smatthewschultz, Chairman and CEO of CleanSpark. “We are evolving into a comprehensive compute platform that is prepared to optimize value from both AI and bitcoin workloads. Our deep expertise in power procurement, infrastructure development, and efficient scaling gives us a unique advantage in meeting surging global demand for compute.” “I’m proud of our results for the fiscal year. Beyond our revenue of $766 million and hashrate growth achievements, we also demonstrated disciplined capital investment and are financially positioned to rapidly become a leading AI infrastructure provider,” said @GaryVec, President and CFO of CleanSpark. “We recently closed a landmark $1.15 billion 0% convertible transaction to accelerate expansion of our power and land portfolio. Our market leading bitcoin mining operations have been supplemented by cash generated from our institutional grade treasury desk. As we continue to execute on our strategies, our goal is to replicate our market leadership across a broader range of compute capabilities.”

Something fascinating is happening that most people are missing. Bitcoin miners are transforming into AI infrastructure providers. Noone missed this but they're not all doing the same thing. I spent time mapping 8 public companies across the value chain. What I found was surprising: there are actually 4-5 distinct business models emerging, each with completely different economics, risk profiles, and upside potential. Let me give you a framework for thinking about this. I used @Agrippa_Inv Classical Value Chain as a base model for this, from his website (link attached in a reply from me after this post and picture attach to this post is also from his site. Thank you for your work). The Core Insight The key question isn't "are they pivoting to AI?" It's "HOW are they monetizing their power infrastructure?" There are actually 4 distinct business models emerging. Each with different economics, risks, and upside potential. Category 3 - The Powered Shell Model Core Scientific $CORZ - 590 MW leased to $CRWZ The capital-efficient play. They provide: Building envelope Power infrastructure Cooling systems The customer funds everything else (~$300M for interiors). $1.5M capex per MW vs $5-10M for competitors. $10.2B contracted over 12 years. Category 5 - The Turnkey Model Cipher $CIFR & TeraWulf $WULF - The landlord model Fully operational data centers, liquid cooling, ready day 1. Tenant brings GPUs. CIFR: $8.5B contracts (AWS, Google/Fluidstack), 3.2 GW pipeline WULF: $14.2B contracts (Fluidstack, Core42), zero-carbon energy Higher capex, higher margins, long-term contracted revenue. Category 5.5 - The Hybrid Iris Energy $IREN - The interesting middle ground $9.7B Microsoft deal. But here's the twist: $IREN owns the GPUs ($5.8B Dell purchase). They're leasing GPU capacity, not data centers. Not quite a CSP. Not quite turnkey. Something in between. Managed GPU hosting. Category 6 - The True Cloud Providers CoreWeave $CRWV & Nebius $NBIS - The platform plays These aren't infrastructure providers. They're running actual cloud platforms. CRWV: $1.9B revenue, 737% YoY growth, Nvidia's golden child NBIS: Former Yandex team, $17.4B Microsoft deal, proprietary full-stack Highest margins. Highest complexity. Some wildcards. Bitfarms $BITF: Most aggressive pivot. Exiting Bitcoin completely by 2027. $814M war chest. Pennsylvania-focused. Soluna $SLNH: Completely different thesis. Co-locates at renewable plants, monetizes curtailed energy. MaestroOS software for grid arbitrage. What This Means The value chain runs from powered shell (30-40% margins) → turnkey (50-60%) → GPU infrastructure (50-60%) → cloud platform (70% ). More value equals more capex equals more risk equals more upside. The key distinction between 5.5 and 6: $IREN owns GPUs and leases capacity. Microsoft gets dedicated infrastructure. $CRWV $NBIS own GPUs and run cloud platforms, customers use APIs and managed services in multi-tenant environments. One is managed hosting. The other is cloud services. Totally different businesses. Why This Matters We're watching a real-time case study in asset transformation. Every company here started as a Bitcoin miner. They all have the same foundational assets. But they're monetizing them completely differently based on their strategy, capital position, and execution capability. The AI infrastructure shortage created massive demand. Miners realized: "We have power infrastructure. That's the actual bottleneck." So they pivoted, but in wildly different directions. The Investment Framework Before you touch any of these, ask five questions: What's their monetization model? (3, 5, 5.5, or 6?) Who owns the GPUs? (Determines capex risk) What's the contract structure? (Revenue visibility) Who are the customers? (Concentration risk) What's the power cost? (Margin sustainability) Pick your risk tolerance: Want proven cloud platform? $CRWV (expensive) or $NBIS (emerging). Want contracted infrastructure revenue? $CIFR, $WULF, $CORZ. Want transformation upside? $IREN, $BITF. Want renewable angle? $WULF, $SLNH. The Risks and Opportunities Real risks: AI bubble pops and demand evaporates. Hyperscalers build their own and disintermediate. Power costs spike and compress margins. GPU refresh cycles accelerate into a capex treadmill. Customer concentration leaves you exposed. Real opportunities: AI infrastructure shortage is genuine. Microsoft, Google, and Amazon are 2-4 years behind demand. These companies can deliver in 12-18 months. And they have the one thing that actually matters: power allocation. That creates pricing power. The Pattern Everyone talks about GPU shortage. The real shortage is power. Training frontier AI models needs 100 MW. Inference at scale needs gigawatts. These companies have secured 5 GW of power capacity. That's the actual scarce resource. Most are 12-24 months from full deployment. The next 18 months determines who executes and who doesn't. The Bottom Line In 5 years, some of these companies will be worth $50B . Others will be restructured or acquired. The difference won't be who had the most megawatts. It'll be who chose the right business model for their assets and execution capability. Strategy > Scale. Business model > Assets. This is capitalism working in real-time. Do the work. Understand the model. Then pick your play.